Ikram Choudhury
Data Scientist & AI/ML Solutions Specialist
Professional Background
- Focus Area: Computer vision, Earth observation, and scalable geospatial models.
- Cefas (Recent Work): Programmed end-to-end MLOps pipelines for terabyte/petabyte-scale marine datasets. Overhauled the UK's Benthic AI litter project, presented at the ICES conference, and implemented RAG, NLP, and DAG models for causal analysis.
- British Airways (Previous Work): Worked as a Data Engineer Consultant, writing production flight analysis code.
Education
- MSc in Space Engineering & BEng in Aerospace Engineering.
- Technical foundation built on satellite radar data and UAV aerodynamics.
Hobbies & Interests
- UI design, 3D animation, and CAD.
- Writing informal blogs on AI, documenting research methods, and solo traveling across the UK and abroad.
Exploring the Site
- AberTech: A suite of three fully functional, commercially viable applications. When unified through a central orchestration layer, they create a comprehensive system designed to assist mountain search and rescue missions and save lives.
- Cloud-DashCam: Ongoing personal work on cloud-based video analysis architectures.
Optimization & Hardware Interoperability
Memory utilization and garbage collection are persistent bottlenecks in Python-based computer vision pipelines. To build cost-effective, hardware-agnostic solutions that don't rely strictly on Nvidia GPUs, I leverage C++ interfaces like the Vulkan API and OpenCL.
A practical example is CompareCompress—a single-day build of an image compression comparer that achieves high-speed performance on legacy hardware (a 2017 MacBook Air) with minimal overhead.